The Multi-Pattern Matching with Online User Requests for Determining Browser Capabilities

Total Page:16

File Type:pdf, Size:1020Kb

The Multi-Pattern Matching with Online User Requests for Determining Browser Capabilities The Multi-Pattern Matching with Online User Requests for Determining Browser Capabilities Ignat Blazhkoa , Tetiana Kovaliukb, Nataliya Kobetsc and Tamara Tielyshevaa a National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, 37, ave. Peremohy, Kyiv, 03056, Ukraine b Taras Shevchenko National University of Kyiv, 64/13, Volodymyrska str., Kyiv, 01601, Ukraine c Borys Grinchenko Kyiv University, 18/2 Bulvarno-Kudriavska str., Kyiv, 04053, Ukraine Abstract One of the most effective ways to get information about Internet users is to analyze User-Agent request header, because it contains information about the user's browser. Quick methods are needed for multi-pattern matching to instances of User Agents. The article considers the algorithm of patterns matching with input strings to determine the capabilities of the browser. The input strings are the headers of User-Agent queries, and strings with special wildcards represent search patterns. The developments in the field of solving the problem of comparing search templates with input lines are considered and analyzed. A new algorithm for solving the problem is proposed. The steps of the algorithm are given and the time complexity of the algorithm for the problem is analyzed. Comparisons of the performance of the received program with analogues are made and the advantage of the offered approach is shown. The program implementation was compared with the most prominent analogues and the advantage of the proposed approach has been shown in terms of execution speed. Keywords 1 Pattern matching, wildcards, browser capabilities, User Agent request 1. Introduction Nowadays, when many companies operate on the Internet, the need to identify user information is very acute. One of the most effective ways to obtain such information is to analyze the User-Agent request HTTP header, which identifies the program making a request to the server and requesting access to the website. The User-Agent request header contains specific information about the hardware and software of the device making the request. This information typically includes information about the browser, web visualization mechanism, operating system, and user’s device, including iPhone, iPad or other mobile device, tablets, desktop computers. Web servers use User-Agent to maintain different web pages in different browsers in case of not adaptive web design, display different content for different operating systems, collect statistics that reflect the browsers and operating systems used by visitors. Web crawling bots also use User-Agent. The data specified in the User-Agent request HTTP header [1] describes browsers (Chrome, Firefox, Internet Explorer, Safari, BlackBerry, Opera), search engines (Google, Google Images, Yahoo), game consoles (PlayStation 3, PlayStation Portable, Wii, Nintendo Wii U), offline browsers (Offline Explorer), links (Link Checker, W3C-check link), electronic readers (Amazon Kindle), validators, cloud platforms, e-mail libraries, scripts. By User-Agent, you can define functions that are supported by a web browser, such as JavaScript, Java Applet, cookie, VBScript, and Microsoft ActiveX. COLINS-2021: 5th International Conference on Computational Linguistics and Intelligent Systems, April 22–23, 2021, Kharkiv, Ukraine EMAIL: [email protected] (I. Blazhko); [email protected] (T. Kovaliuk); [email protected] (N. Kobets), [email protected] (T. Tielysheva) ORCID: 0000-0002-1383-1589 (T. Kovaliuk); 0000-0003-4266-9741 (N. Kobets); 0000-0001-5254-3371 (T. Tielysheva) ©� 2021 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). CEUR Workshop Proceedings (CEUR-WS.org) The User-Agent string is used to match the content when the source server selects the appropriate content or operating parameters to respond to the received user request. Thus, the concept of adapting content to the needs of users is realized. The User-Agent string is one of the criteria by which search engines can be excluded from accessing certain parts of the website using the robot exclusion standard (according to the robots.txt file). Information about users has obvious applications in the advertising sector, namely in cases where the user's device can operate as a criterion for determining the user's affiliation to the target market. Web analysis is an equally important use case. Due to the availability of information about customers, further analysis can significantly improve the decision about the published content; increase the conversion of the site by turning site visitors into real customers. Analyzing User-Agent request headers that are constantly updated also means that businesses will be aware of the emergence of new devices that visit their resource, and therefore related problems will be identified at an early stage. However, parsing the User-Agent request header is not a trivial task. Many software libraries are used to solve this problem by balancing between accuracy and speed of obtaining results. There are services that provide definition and analysis of the User-Agent request header, in particular, deviceatlas.com. The accuracy problem can be solved by using open lists of User-Agent templates, which are constantly updated. There are more than 62 million User-Agent strings [1], so existing solutions cannot quickly work with all of them. The speed of response to the client is another important indicator for business. Therefore, quick methods of analyzing these patterns are needed. These are methods for matching a large number of search patterns to User-Agent instances. The development of algorithms based on which libraries will be created to determine the capabilities of browsers is an urgent and necessary task for any web service. The purpose of the research is to develop an algorithm for multi-pattern matching with wildcards to be able to detect browser capabilities based on browscap.org data [2]. The library should work faster than its alternatives − the speed should be comparable to libraries that use simplified (inaccurate) methods to detect browser capabilities. Browscap is an open project dedicated to collecting and disseminating a database of browser capabilities, actually a set of different files describing browser capabilities. Browscap has built-in support for using these files. To achieve this goal we need to solve the following tasks: • figure out browscap.org data structure and similarities between different user agents of the same family; • review and analyze existing studies that solve the problem of multi-pattern matching with wildcards • develop an algorithm for multi-pattern matching with the best known lower bound time complexity for determining the browser capabilities; • develop software implementation of the algorithm and compare the speed of the proposed algorithm with analogue libraries, which give approximate results. 2. Problem Statement Let the set of search patterns P = {p1, p2 ,..., pm } be input to the search engine. Each search pattern pi , i =1, m is a string of characters that belong to the alphabet Σ . Each pattern contains special characters "?" and "*", that do not belong to this alphabet, and are called wildcards. The character "?" corresponds to any single character in the alphabet Σ , the character "*" corresponds to any string of characters in the alphabet Σ , including the empty string. The search pattern pi , i =1, m can be represented as a string pi = ci1ki1ci2ki2...cij kij *, where cij is the j -th wildcard character for the i -th pattern, kij is a keyword from a dictionary of unique keywords W , that is, a string of characters in the alphabet Σ . There is an additional condition for the wildcard character ci1 : it can be either skipped or equal to the character "*". Input requests are in the form of a string S ={si },i =1, n , consisting of n characters in the alphabet Σ . It is necessary for each input string S ={si },i =1, n to find all the search patterns from the set P that matches to the query string, i.e. get a set of patterns A ={p , p ,..., p }, where p denotes a1 a2 aanslen aS the search pattern for the S -th request. The number of requests is not limited. Requests must be answered as they are received. One of the features of this task is that the set of search patterns is very large and is in the millions of copies. Query string sizes typically range from 100 to 200 characters. 3. Related Works The main tasks of text string analysis that arise in the development of information verification tools in information retrieval systems include: • the task of matching text strings; • the problem of calculating the distance between text strings; • the problem of fuzzy match text strings; • the task of finding the longest repeating text substring. The algorithms for text string analysis are presented in the works of domestic and foreign researchers, in particular Stephen G.A., Knuth D.E., Morris J.H., Pratt V.R., Karp R.M., Rabin, M.O., Boyer R.S., Moore J.S., Fischer M.J., Hirschberg D.S., Hunt J.W., Szymanski T.G., Landau G.M., Vishkin U. Hamming R.W., Levenshtein V.I. The task of determining the browser capabilities based on the analysis of the User-Agent HTTP header is reduced to the task of matching text strings, which are used as search patterns and incoming requests. Algorithms for matching search patterns with input strings have many important applications and are used in antivirus software, systems for detecting various types of attacks and intrusion prevention [3], in-text compression, text search, data mining, programming grammar checking rules [4] etc. For example, different types of attacks can be defined using rules, which are regular expressions that match a set of possible strings. Algorithms for working with regular expressions require pre- processing which is very memory-intensive and time-consuming. An example of the application of the algorithm for matching search patterns with input strings is the study of DNA sequences, where wildcards are used as a replacement for some components of protein sequences.
Recommended publications
  • The Problem of Fuzzy Duplicate Detection of Large Texts
    The problem of fuzzy duplicate detection of large texts E V Sharapova1 and R V Sharapov1 1Vladimir State University, Orlovskaya street 23, Murom, Russia, 602264 Abstract. In the paper, considered the problem of fuzzy duplicate detection. There are given the basic approaches to detection of text duplicates–distance between strings, fuzzy search algorithms without indexing data, fuzzy search algorithms with indexing data. Thereview of existing methods for the fuzzy duplicate detection is given. The algorithm of fuzzy duplicate detection is present. The algorithm of fuzzy duplicate texts detection was implemented in the system AVTOR.NET. The paper presents the results of testing of the system.The use of filtering text, stemming and character replacement, allow the algorithm to found duplicates even in minor modified texts. 1. Introduction Today computer technology and the World Wide Web have become part of our lives. Billions of users use the Internet every day. The Internet contains a huge number of documents. Many books are transformed in a digital form. Users can read books, articles, newspapers directly from computers and electronic gadgets. Today, in the Internet there are many studying materials: lectures, textbooks, methodical literature, etc. There are big collections of essays, course and diploma projects and scientific dissertations. It leads to problem un-controlling coping of information. The text of document may be a part of another document, formed from several documents, may be modified part of another document (with a change of words to synonyms, endings, times, etc.). In this way, the document is fuzzy duplicate of other documents. There is the problem of text checking to presence of fussy duplicates.
    [Show full text]
  • Algorithms for Approximate String Matching Petro Protsyk (28 October 2016) Agenda
    Algorithms for approximate string matching Petro Protsyk (28 October 2016) Agenda • About ZyLAB • Overview • Algorithms • Brute-force recursive Algorithm • Wagner and Fischer Algorithm • Algorithms based on Automaton • Bitap • Q&A About ZyLAB ZyLAB is a software company headquartered in Amsterdam. In 1983 ZyLAB was the first company providing a full-text search program for MS-DOS called ZyINDEX In 1991 ZyLAB released ZyIMAGE software bundle that included a fuzzy string search algorithm to overcome scanning and OCR errors. Currently ZyLAB develops eDiscovery and Information Governance solutions for On Premises, SaaS and Cloud installations. ZyLAB continues to develop new versions of its proprietary full-text search engine. 3 About ZyLAB • We develop for Windows environments only, including Azure • Software developed in C++, .Net, SQL • Typical clients are from law enforcement, intelligence, fraud investigators, law firms, regulators etc. • FTC (Federal Trade Commission), est. >50 TB (peak 2TB per day) • US White House (all email of presidential administration), est. >20 TB • Dutch National Police, est. >20 TB 4 ZyLAB Search Engine Full-text search engine • Boolean and Proximity search operators • Support for Wildcards, Regular Expressions and Fuzzy search • Support for numeric and date searches • Search in document fields • Near duplicate search 5 ZyLAB Search Engine • Highly parallel indexing and searching • Can index Terabytes of data and millions of documents • Can be distributed • Supports 100+ languages, Unicode characters 6 Overview Approximate string matching or fuzzy search is the technique of finding strings in text or dictionary that match given pattern approximately with respect to chosen edit distance. The edit distance is the number of primitive operations necessary to convert the string into an exact match.
    [Show full text]
  • A Parallel String Matching Tool
    Clgrep: A Parallel String Matching Tool Peng Wu a thesis submitted for the degree of Master of Science at the University of Otago, Dunedin, New Zealand. March 18, 2013 Abstract In this study, we widely investigate the problem of string matching in the context of Heterogeneous Parallel Computing. A overview of string match- ing is made, in which the different forms of string matching problem are distinguished, and the classifications of string matching algorithm are dis- cussed. As an alternative to grep for computational intensive string matching and in addition to support the research of the study, a parallel exact string matching utility `Clgrep' is developed. By experimental studies, we investigate the use of heuristics-based algo- rithms, specifically QS and Horspool algorithms, in the context of Het- erogeneous Parallel Computing. The results suggest that the performance of Heterogeneous Parallel Computing matching, either on multi-core CPU or GPU, is highly related to the computational intensity of certain cases. When computational power is intensively required, the SIMD Parallel Com- puting model of Clgrep can be several times more efficient than correspond- ing sequential matching program. ii Acknowledgements This study has been a difficult task as the problem is novel and involves many programming challenges. Everyday, I am fighting with the errors in the program and the issues raised in the experiment. I do not remember how many hours were spent in the complex experiment. However, I can not forget the people helped me during the study. I would like to thank my parents, without them my study would merely be a dream.
    [Show full text]
  • Improvising of Web Recommendation System by Using K-Means and Bitap Algorithm
    November 2017, Volume 4, Issue 11 JETIR (ISSN-2349-5162) IMPROVISING OF WEB RECOMMENDATION SYSTEM BY USING K-MEANS AND BITAP ALGORITHM 1 Dipali Wankhede,2 S.G.Tuppad 1ME Student,2Professor 1Computer Science, 1MSSCET, Jalna,India Abstract— The increase in the amount of information over the Internet in recent years has led to the risk of data flooding, which has resulted in problems accessing user-related information. In addition, the increase in the number of websites and web pages has made it difficult for webmasters to determine the content of their content. The online user's information needs can be obtained by evaluating the user's web usage. Web Based Data Mining (WUM) is used to extract knowledge from web users. Access logs are accessed by using data mining techniques. One of the WUM applications is a guideline, a privacy filtering technique used to determine whether a user approves a specified item or identifies a list of things that may be important to the user. This article is an architecture that combines product data with user access log data and generates a set of instructions for that user. The operation records interesting results in terms of precision, recall, and metrics. Index Terms— Web Usage mining, Online Web Recommendation System, Clustering, Pattern Matching, Boyer Moore, K- means, Recommendation I. INTRODUCTION In recent years, eCommerce, web services and web-based information systems have been used rapidly. The explosion of the web and the emergence of ecommerce have encouraged designers to develop. The ecommerce product introduction system has transformed the global business landscape.
    [Show full text]
  • Java Exception Matching in Real Time Using Fuzzy Logic Master of Science Thesis
    Java exception matching in real time using fuzzy logic Master of Science Thesis Karl Tillström Chalmers University of Technology University of Gothenburg Department of Computer Science and Engineering Göteborg, Sweden, May 2010 The Author grants to Chalmers University of Technology and University of Gothenburg the non-exclusive right to publish the Work electronically and in a non-commercial purpose make it accessible on the Internet. The Author warrants that he/she is the author to the Work, and warrants that the Work does not contain text, pictures or other material that violates copyright law. The Author shall, when transferring the rights of the Work to a third party (for example a publisher or a company), acknowledge the third party about this agreement. If the Author has signed a copyright agreement with a third party regarding the Work, the Author warrants hereby that he/she has obtained any necessary permission from this third party to let Chalmers University of Technology and University of Gothenburg store the Work electronically and make it accessible on the Internet. Java exception matching in real time using fuzzy logic KARL TILLSTRÖM © KARL TILLSTRÖM, MAY 2010. Examiner: BROR BJERNER Chalmers University of Technology University of Gothenburg Department of Computer Science and Engineering SE­412 96 Göteborg Sweden Telephone + 46 (0)31­772 1000 Department of Computer Science and Engineering Göteborg, Sweden May 2010 Abstract This thesis deals with the problem of matching Java exceptions that originates from the same problem, but looks a bit dissimilar. The match- ing must be done within reasonable time to be able to be carried out automatically.
    [Show full text]
  • Needleman–Wunsch Algorithm
    Not logged in Talk Contributions Create account Log in Article Talk Read Edit View history Search Wikipedia Wiki Loves Love: Documenting festivals and celebrations of love on Commons. Help Wikimedia and win prizes by sending photos. Main page Contents Featured content Current events Needleman–Wunsch algorithm Random article From Wikipedia, the free encyclopedia Donate to Wikipedia Wikipedia store This article may be too technical for most readers to understand. Please help improve it Interaction to make it understandable to non-experts, without removing the technical details. (September 2013) (Learn how and when to remove this template message) Help About Wikipedia The Needleman–Wunsch algorithm is an algorithm used in bioinformatics to align Class Sequence Community portal protein or nucleotide sequences. It was one of the first applications of dynamic alignment Recent changes Worst-case performance Contact page programming to compare biological sequences. The algorithm was developed by Saul B. Needleman and Christian D. Wunsch and published in 1970.[1] The Worst-case space complexity Tools algorithm essentially divides a large problem (e.g. the full sequence) into a series of What links here smaller problems and uses the solutions to the smaller problems to reconstruct a solution to the larger problem.[2] It is also Related changes sometimes referred to as the optimal matching algorithm and the global alignment technique. The Needleman–Wunsch algorithm Upload file is still widely used for optimal global alignment, particularly when
    [Show full text]
  • The Evaluation of Text String Matching Algorithms As an Aid to Image Search
    JOURNAL OF APPLIED COMPUTER SCIENCE Vol. 26 No. 1 (2018), pp. 33-62 The Evaluation of Text String Matching Algorithms as an Aid to Image Search Joanna Ochelska-Mierzejewska1 1Lodz University of Technology Institute of Information Technology ul. Wólcza´nska215, 90-924 Łód´z,Poland [email protected] Abstract. The main goal of this paper is to analyse intelligent text string matching methods (like fuzzy sets and relations) and evaluate their useful- ness for image search. The present study examines the ability of different algorithms to handle multi-word and multi-sentence queries. Eight different similarity measures (N-gram, Levenshtein distance, Jaro coefficient, Dice coefficient, Overlap coefficient, Euclidean distance, Cosine similarity and Jaccard similarity) are employed to analyse the algorithms in terms of time complexity and accuracy of results. The outcomes are used to develop a hi- erarchy of methods, illustrating their usefulness to image search. The search response time increases significantly in the case of data sets containing sev- eral thousand images. The findings indicate that the analysed algorithms do not fulfil the response-time requirements of professional applications. Due to its limitations, the proposed system should be considered only as an illustra- tion of a novel solution with further development perspectives. The use of Polish as the language of experiments affects the accuracy of measures. This limitation seems to be easy to overcome in the case of languages with simpler grammar rules (e.g. English). Keywords: text comparison, N-gram, Levenshtein distance, Jaro coefficient, Dice’s coefficient, Overlap coefficient, Euclidean distance, Cosine similarity, Jaccard similarity.
    [Show full text]
  • Accelerating Approximate Pattern Matching with Processing-In-Memory (PIM) and Single-Instruction Multiple-Data (SIMD) Programming
    Accelerating Approximate Pattern Matching with Processing-In-Memory (PIM) and Single-Instruction Multiple-Data (SIMD) Programming Damla Senol Cali1 Zülal Bingöl2, Jeremie S. Kim1,3, Rachata Ausavarungnirun1, Saugata Ghose1, Can Alkan2 and Onur Mutlu3,1 1 2 3 Poster SEQ-15 Accelerating Approximate Pattern Matching with Processing-In-Memory (PIM) Problem: and Single-Instruction Multiple-Data (SIMD) Programming Damla Senol Cali1, Zülal Bingöl2, Jeremie S. Kim1,3, Rachata Ausavarungnirun1, Saugata Ghose1, o Bitap algorithm can perform string matching with Can Alkan2 and Onur Mutlu3,1 1 Carnegie Mellon University, Pittsburgh, PA, USA 2 Bilkent University, Ankara, Turkey 3 ETH Zürich, Zürich, Switzerland fast and simple bitwise operations. Bitap Algorithm Problem & Our Goal Processing-in-Memory TSV HBM DRAM Die Bitap algorithm (i.e., Shift-Or algorithm, or Baeza- Problem: Microbump o Due to the von Neumann architecture, memory Yates-Gonnet algorithm) [1] can perform exact string o The operations used during bitap can be performed in matching with fast and simple bitwise operations. parallel, but high-throughput parallel bitap 3D-Stacked DRAM Wu and Manber extended the algorithm [2] in order computation requires a large amount of memory to perform approximate string matching. Processor (GPU/CPU/SoC) Die bandwidth that is currently unavailable to the PHY Logic Die PHY . bus between the CPU and the memory is the § Step 1 – Preprocessing: For each character in the processor. Interposer alphabet (i.e., A,C,G,T), generate a pattern o Read mapping is an application of approximate string Package Substrate bitmask that stores information about the matching problem, and thus can benefit from existing o Recent technology that tightly couples memory presence of the corresponding character in the techniques used to optimize general-purpose string and logic vertically with very high bandwidth bottleneck of the high-throughput parallel bitap pattern.
    [Show full text]
  • Real Time Pattern Based Melodic Query for Music Continuation System
    Real time Pattern Based Melodic Query for Music Continuation System Sanjay Majumder Benjamin D. Smith Indiana University - Purdue University Indianapolis Indiana University - Purdue University Indianapolis Indianapolis, Indiana, USA Indianapolis, Indiana, USA [email protected] [email protected] ABSTRACT 1 INTRODUCTION This paper presents a music continuation system using pattern The MIDI file format is one of the most popular and broadly applied matching to find patterns within a library of MIDI files using areal- forms of symbolic music representation. Despite having limitations, time algorithm to build a system which can be used as interactive its wide adoption has resulted in many, many MIDI databases that DJ system. This paper also looks at the influence of different kinds are available on the Internet. With the rapid advancement of Digital of pattern matching on MIDI file analysis. Many pattern-matching Audio Workstations and the advent of inexpensive MIDI controllers algorithms have been developed for text analysis, voice recognition and keyboards, musicians can easily compose, distribute and share and Bio-informatics but as the domain knowledge and nature of MIDI files. the problems are different these algorithms are not ideally suitable Pattern matching is a primitive problem in the field of computer for real time MIDI processing for interactive music continuation science as well as in information retrieval. With the advancement system. By taking patterns in real-time, via MIDI keyboard, the sys- of computer science, researchers have developed many pattern- tem searches patterns within a corpus of MIDI files and continues matching algorithms, but most of these algorithms are problem playing from the user's musical input.
    [Show full text]
  • An Exploration of the Efficiency of String Matching Algorithms
    INTERNATIONAL JOURNAL OF RESEARCH CULTURE SOCIETY ISSN: 2456-6683 Volume - 2, Issue - 3, Mar – 2018 UGC Approved Monthly, Peer-Reviewed, Refereed, Indexed Journal Impact Factor: 3.449 Publication Date: 31/03/2018 An Exploration of the Efficiency of String Matching Algorithms Vennila Santhanam, Assistant Professor, Department of Computer Science, Auxilium College (Autonomous), Vellore. Tamil Nadu, India. Email - [email protected] Abstract: String searching is a process that refers to the problem of finding all the occurrences of a particular string in the given text. String matching algorithms plays a vital role in finding such appropriate match in limited duration. String matching algorithms are used in many areas such as Spell Checkers, Spam Filters, Intrusion Detection System, Search Engines, Plagiarism Detection, Bioinformatics, Digital Forensics and Information Retrieval Systems, Computational molecular biology. The World Wide Web provides settings in which efficient string matching algorithms are necessary. In this study the researchers review most of the String matching algorithms. It is a class of string algorithms that tries to find the occurrences and location of one or several strings (also called as patterns) are found within a larger string or text. Key Words: String matching, Patterns, Occurrences. 1. INTRODUCTION: Let Σ be an alphabet set. The most basic example of string searching is where both the pattern and searched text are sequence of elements of Σ. Kinds of String searching String searching may involve a very long string, called the Haystack or it may be a very short string, called the Needle. The ultimate goal of string searching is to find one or more occurrences of the needle string existing within the haystack.
    [Show full text]
  • Detecting Near Duplicates in Software Documentation
    Detecting Near Duplicates in Software Documentation D.V. Luciv1, D.V. Koznov1, G.A. Chernishev1, A.N. Terekhov1 1 Saint Petersburg State University 199034 St. Petersburg, Universitetskaya Emb., 7–9 E-mail: {d.lutsiv, d.koznov, g.chernyshev, a.terekhov}@spbu.ru 2017-11-15 Abstract Contemporary software documentation is as complicated as the software itself. During its lifecycle, the documentation accumulates a lot of “near duplicate” fragments, i.e. chunks of text that were copied from a single source and were later modified in different ways. Such near duplicates decrease documentation quality and thus hamper its further utilization. At the same time, they are hard to detect manually due to their fuzzy nature. In this paper we give a formal definition of near duplicates and present an algorithm for their detec- tion in software documents. This algorithm is based on the exact software clone detection approach: the software clone detection tool Clone Miner was adapted to detect exact dupli- cates in documents. Then, our algorithm uses these exact duplicates to construct near ones. We evaluate the proposed algorithm using the documentation of 19 open source and com- mercial projects. Our evaluation is very comprehensive — it covers various documentation types: design and requirement specifications, programming guides and API documenta- tion, user manuals. Overall, the evaluation shows that all kinds of software documentation contain a significant number of both exact and near duplicates. Next, we report onthe performed manual analysis of the detected near duplicates for the Linux Kernel Documen- tation. We present both quantative and qualitative results of this analysis, demonstrate algorithm strengths and weaknesses, and discuss the benefits of duplicate management in software documents.
    [Show full text]
  • The Assignment Problem on Sets of Strings
    InternalReport2012–17 August2012 Universiteit Leiden Opleiding Informatica The Assignment Problem on Sets of Strings Thijs van Ommen BACHELOR THESIS Leiden Institute of Advanced Computer Science (LIACS) Leiden University Niels Bohrweg 1 2333 CA Leiden The Netherlands Contents 1 Introduction 1 2 Approximate string matching 2 2.1 String distance measures . 2 2.1.1 Edit distance and variants . 2 2.1.2 Token-based distances . 3 2.1.3 Block edit distances . 4 2.1.4 Specialized distances . 6 2.1.5 Compression-based distances . 6 2.2 Approximate string matching algorithms . 7 2.2.1 Online algorithms . 8 2.2.2 Algorithms using preprocessing . 8 2.3 Other problems involving approximate string matching . 9 2.3.1 File comparison . 9 2.3.2 Record linkage . 10 2.3.3 Inference . 11 3 The assignment problem and the Hungarian algorithm 11 3.1 Theory and basic form of the Hungarian algorithm . 12 3.2 Improvements of the algorithm . 13 4 Our algorithm 15 4.1 Distance measure . 15 4.2 Hashing . 18 4.3 Computing lower bounds on distances . 19 4.4 Computing distances exactly . 20 4.5 Adapting the Hungarian algorithm . 22 4.5.1 Finding the minimum . 22 4.5.2 Labels . 23 4.5.3 Creating the row queues . 24 4.5.4 Rectangular matrices . 25 4.5.5 Voluntarily assigning fewer lines . 25 4.6 Theoretical performance analysis . 26 5 Experiments 27 5.1 Correctness . 27 5.2 Speed . 27 6 Conclusion 30 References 31 i ii 1 Introduction In this thesis, we will develop a tool that compares two text files to get an overview of their differences, in a way similar to what the diff tool does.
    [Show full text]